Tests of the money-output relation using disaggregated data

Either anticipated or unanticipated money affects output in fourteen of twenty U.S. manufacturing industries. In most of these instances, however, Akaike's final prediction error criterion indicates that money enters an industry's output equation with lags of three months or less. For just two industries, tobacco manufacturing and textile mill products, are there clear indications that money is not neutral at extended lags. Each of these industries is concentrated in one or two states suggesting that monetary policy may affect output through a regional credit channel.